Skip to main content
U.S. flag

An official website of the United States government

Back to Blog

Human-centered design for IT centralization, part 2 - Deciding whether or not to centralize

This is Part 2 in a series of posts on the importance of human-centered design when evaluating IT centralization. As part of a 10x project, we’ve synthesized 18F’s learnings from agency partners who’ve been through centralization efforts before and have wisdom to share. The series explores how and why taking the time to prioritize users will mitigate risks and yield services that work better for the people they serve. In this post, we’ll help you answer the “to centralize or not to centralize” question.

Many agencies trying to improve their internal IT strategy wonder if centralization is the answer. While there might be opportunities to consolidate, streamline, and reorganize common IT functions across departments, there may also be concerns:

  • Does it make sense to reorganize and streamline some IT services but not others?
  • Will this effort reduce costs and increase efficiency or would it be a budget drain?
  • Will it help employees or be a burden on them?

After interviewing several folks who’ve been through a centralization initiative, we’ve developed some best practices for navigating these questions in a user-centered way:

  1. Set specific, realistic, measurable goals
  2. Bring the right people to the table
  3. Collaborate to identify the problem and set the vision
  4. Avoid one-size-fits-all
  5. Anticipate the hurdles

Most importantly, incorporating human-centered design into your centralization evaluation should be part of your strategy. Because at the end of the day, if your users are not benefiting from the centralization, it’s probably not a good idea.

1. Set specific, realistic, measurable goals

The President’s Management Agenda encourages agencies to improve transparency, customer service, service sharing, and reduce redundancies in service of the agency’s mission. These are all really big goals. As you move to implement them, it will be important to prioritize one. This doesn’t mean not achieving the others, but choosing one to drive toward first.

Your goals should depend on what type of function you’re looking to centralize. Whether centralizing is appropriate or not often depends on how often a service repeats across an agency and how important flexibility is to deliver on the agency’s mission.

Below is a spectrum of several common services. On the left are services that lend well to centralization. Usually people don’t need a lot of flexibility in how these services are delivered — they just need to work. On the right are services that don’t lend themselves well to centralization. For example, people often need more flexibility in how they develop software to deliver most effectively.

Chart showing a list of services from centralization friendly to it allows flexibility

Spend some time defining how centralizing the IT function you have in mind would help your users achieve their goals and be more agile, cost effective, and compliant. Try and map goals based on the effort they will take to achieve and the impact for users. Prioritize high positive impact and medium-to-low effort.

Clarifying your goals will help with communication across your organization and identify ways other folks in the agency can play a role.

2. Bring the right people to the table

Nothing is worse than investing large amounts of money and time into something that nobody likes or uses. This scenario is often the result of not having the right people at the table when making decisions about what to build or buy. Those missing voices will help challenge assumptions around what is useful.

In the user-centered design world, we believe the people who use services everyday should always be at the table. At the beginning of your effort to understand whether to centralize an IT function or not, identify front-line partners who use the services.

Front-line partners can help you understand how the services are used, the problems users have, and the opportunities users see. You will be most successful if you work together.

Front-line managers, for example, often have the best insight into why current systems are set up the way they are and what features they need in the future. They’re experts in their areas and should have decision power over what services are delivered. IT managers should have decision power over how they’re delivered.

3. Organize collaborative exercises to set the vision

Here’s a helpful guide for organizing exercises with your team that will help you evaluate a centralization initiative, or jump to a specific exercise below to learn more.

1. Facilitate a consensus-building workshop

2. Capture hopes and fears

3. Define the current state

4. Get a deeper understanding of existing services

5. Send an open-ended survey to users

6. What are good candidates for consolidation?

7. Dig deeper into “good” and “bad” candidates

8. Define success metrics

Don’t get too attached to a detailed plan or set vision at the early stages of a project. Remember: you will learn as you go and may need to adjust your plans along the way. If you’re working in an agile way, this is natural. As Graham Kenny writes, “strategic plans are less important than strategic planning.”

4. Avoid one-size-fits-all

It can be tempting to solve everyone’s problems in one big bang, but one size never fits all when it comes to centralization. Users have unique needs that one-size-fits-all solutions may meet awkwardly at best. Services that don’t fit well may, at best, get left on the shelf and, at worse, cost more money to customize or fix in the future.

Reduce risk by starting small

In addition to meeting user needs, the numbers show that smaller projects are less risky than larger ones. According to the Standish Group Chaos Report, small projects succeed 62 percent of the time, compared to 6 percent of large projects. Large projects are naturally more complex. More moving parts and larger teams are harder to manage. This is a large risk to take when dealing with a centralization effort that will impact many people.

If you’re trying to address many different problems, consider restructuring the work into smaller chunks to avoid risk and complexity. 18F’s acquisitions team calls this approach modular contracting.

Give users a choice

A centralized service offering will be better received if users feel like they have a choice. Mandating that users use specific tools is more likely to lead to users developing work arounds and using shadow IT. You can develop and maintain trust by making it voluntary and low-to-no cost to use your centralized offering. The competition with other tools will also help you focus on delivering more value to the users than other options.

Keep track of which tools people are using and identify which tools/services are most popular among users. This practice will help you understand how your users are currently solving problems. It’s also a fairly lightweight way to allow users to drive IT investments.

5. Anticipate the hurdles

Don’t wait until the end of a project to make the following considerations:

Start early on getting an Authority to Operate (ATO)

Leaving this process until the end will only add time to your project, delaying value to your users. Make sure you involve your policy and security team members early to understand compliance, which will make the end of the project go smoother.

Consider your rollout plan

Identifying what IT functionality should exist is only the beginning. Start thinking about how you will communicate the changes you make to your team, how you will phase out or update existing systems, and how you will maintain trust with your users.

At the end of the day, centralization may or may not be the right strategy for your agency. Approaching the decision in a user-centered way will help you gain trust with your team and reduce the risk of making a poor investment.

Next in our series, we’ll talk about working with vendors and how to build or buy a centralized solution from them in a user-centered way.

Related posts